- Email: [email protected]
The wearing power of steel rails in relation to their chemical composition and physical properties
Mar., 1881.]
Dudley--Steel Rails.
177
tile combustible portion Of "Nixon's navigation coal" was intrinsically /12"7519--12"0906 X 1 0 0 = ~ 5"47 per centum superior in water \ ! 12"0906 vaporizative power to the pound of the combustible portion of the Cumberland semi-bituminous coal. That is to say, if both coals had the same percentage and kind of refuse such would be their relative value in, water heating effect. This difference may very easily be allowed. The very great difference in the water vaporizative power of equal weights of the two crude coals prevents their use for measuring the economic performances of themachinery in the two cases.
THE W E A R I N G P O W E R O F S T E E L R A I L S I N R E L A T I O N TO T H E I R C H E M I C A L C O M P O S I T I O N A N D PHYSICAL PROPERTIES.* By CHARLES B. D U D L E Y , Ph.D., Chemist Pennsylvania Railroad Company, Altoona, Pa. Read at the Philadelphia Meeting of the American Institute of Mining Engineers, held at the Franklin Institute, February 17, 1881.
TREO. N. ELY, Esq., Superintendent of Motive Power, Pennsylvania Railroad. DEAR S,R--It is now nearly three years since my first report to you on the subject of steel rails was written. That report, as you will At the Lake George meeting of the Institute in October, 1878, I had the honor of presenting to the [nstitute, through the kind permission of the officers of the Pennsylvania Railroad Company, the results of a study of twenty-five examples of steel rails, which had all been in actual service. Considerable discussion followed tile publieation of that paper, and there seemed to be a strong disinclination, especially on the part of the steel rail manufacturers, to accept the conclusions presented. One of the principal objections urged against the conclusions drawn was that they were based on too few samples; in other words, that no conclusions safe to act upon could be drawn from the examination of twenty-five rails. In view of this criticism it wm~ decided to repeat the investigation with a larger number of samples, and with the aid of the experience gained in the first investigation. The results of this Second study of steel rails are, by the permission of the officers of the Pennsylvania :Railroad Company, herewith presented to the Institute, with the sincere desire that they may aid in adding to our knowledge of this most important product. Like the previous paper, this is in the form of a report to one of the officers of the company, which will account lbr the style, and the manner of presenting tile data. WHOLE ~O. VOL. CXI.--(TtHRD SERIES, Vol. lxxxi.) 12
178
Dudley--Steel Rails.
[Jour. Frank. Inst,
remember, dealt principally with the question of the relation between the chemical composition and physical properties of steel rails and :their power to resist crushing and fracture in actual service. Other matters were referred to or touched upon in that report, but the main question was, Why do some rails crush or break in service while ,others do not ? You will doubtless remember that the principal eon,clusion arrived at was, that the softer rails are less liable to crush or break in service than the hard ones. Or, in other words, so far as ~onclusions could be drawn fi'oln the chemical analysis and physical ~est of 25 samples of steel rails which had actually been in service, these conclusions were that those rails which have the smaller amounts of carbon, plmsphorus, silicon and manganese are less liable to crush ,or break in service than those which have larger amounts of these ,elements. Or, again, looked at in the light of physical tests, those rails which have the lower tensile strength and the greater elongation are the ones which give the least trouble from breaking or crushing in ~rack. In the report just referre~l to, the question of the wearing power of steel rails was not made prominent, and from that report no positive :and definite information could be obtained as to what quality of steel would give rails that would endure the greatest amount of traffic with £he least loss of metal. And yet, of the three principal causes which ~ccasion the removal of rails from the track, viz., broken, crushed and worn out, perhaps the latter is of the most importance'. With ~he improvement in maintenance of way which has characterized the Pennsylvania Railroad during the last five or six years, the removal .of rails from track from the first two of these causes has, if I am right, quite notably diluinshed. This certainly is true with regard to broken rails. And if, as time advances, the number of crushed rails ,shall dhninlsh, both because of the continued improvement in main~euance of way, before referred to, and because, owing to improved and better methods at the steel works, there are fewer crushed rails caused by physical defects in the steel, the question of the wearing power of steel rails ohviously becomes the all-important one. In view of these considerations, it was thought that an investigation into the relation between the wearing power of steel rails and their chemical composition and physical properties could not fail to throw light upon a question of vital importance in the nmnagement of the Pennsylvania Railroad. - The results of such an investigation are presented in the
Mar., 1881.]
Dudley--Steel Rails.
179
ibllowing report, which deals not, as did the previous report, with the relations between the chemical and physical characteristics of steel rails and their power to resist crushing or fracture, but entirely with the relation between these characteristics and wear. On ally railroad the rails are or may be called upon to perform their service under quite varying conditions. Oil a railroad like the Pennsylvani'a Railroad, for example, there are levels and grades, and there are tangents and curves, and tllcre are combinations of the levels and grades with the tangents and curves. Moreover, the rails oll the high sides of curves do their service under different conditions from those on the low sides of curves. So that, as far as kind of service is concerne(i, a rail may be called upon to do its work under one of these six conditions, viz., on level tangent, on high side of level curve, oil low side of level curve, on grade tangent, on high side of grade curve, or on low side of grade curve. Any investigation, therefore,into the wearing power of steel rails, which shall be of service in determining what rails are best for the whole road, must take into account these slx conditions. O f course there arc many different grades and curves of differeut radius on the Pennsylvania Railroad ; but to study wear on each grade, and for every radius of curvature, would make a problem almost liih-long. In order, therefore, to make the work manageable it becomes necessary to choose some average grade and some medium degree of curvature, as representing the dittbrent grades and curves on the road. This has been done in the work about to be described. Now, the problem before us is: What chemical composition and w~mt physical properties are characteristic of those rafts which in actual service have lost least metal in i,roportiou to the tonnage that has passed over them? In order to get the information neces~ry to answer this question, 64 rails were taken from the track in July, 1879, and subjected to chenfical analysls and physical test, as is detailed ihrthe!" Oil. Sixteen of these rails were taken from level tangents and 16 from level curves, 8 from the high side and 8 from the low side of the curves. Again, 16 rails were taken from grade tangents and 16 from grade curves, 8 from the high side arid 8 from the low side of these cm'ves. The 32 rails on grades were all taken between Conemaugh and Altooua, and the 32 on levels beCween Tyrone and Miftlin. Furthermore, 32 of the rails were taken fl'om the north track, and 32 from the south track. The principle governing the selection was to
180
Dudley--Steel Rails.
[Jour. Frank. Inst..
secure 8 as slow-wearing rails and 8 as rapid-wearing rails as could be found on each of the fimr conditions of service--level tangents, level curves, grade tangents and grade curves--the rails on curves bein~ taken, as has been already described, one-half from the high side or' the curve and one-half from the low side of the curve. In the actual selection a pair of callipers was used, and the loss in height of a rail, compared with it~ time of service, gave a su~eiently accurate measure of the rate of wear of a rail to determine its selection. It should be stated here that since the rails were selected in July, after the usual annual removal from the track of worn-out rails, which takes place earlier in the season, the more rapid-wearing rails obtained for examination are not as marked examples of rapid-wearlng rails as would have been obtained had the selection been earlier made. With regard to the 64 rails chosen, it may also be stated that not one of them had broken in service, and only one showed any signs of crushing ; so that, as far as every quality is concerned, except their rate of wear, ali of them might be classed as good rails. It was thought that a chemica[examination and physical test of the 64 rails, compared with the loss of metal they had suffered by the tonnage which had pa;sed over them, could not Sail to throw some light on the qnestion as to what kind of steel in rails gives best wear. These rails having been selected were removed from the track and sent to Altoona. They were then cleaned and weighed on an ordinary platform scale, their length measured as accurately as possible with a steel tape, and their height callipered near both ends and in the middie. The weight of the whole rail divided by its length obviously gives the present weight per yard of these worn rails. The weights obtainltl in this manner, however, were not subsequently used, as will be explained further on. After the weighing, five feet were cut off from the end of each rail for test purposes. The test-pieces, except those used in the bending test, were all cut from the head of the rail. Two pieces for tensile test, two for torsion test, and two pieces from the web for bending test, as well as a section of the rail ½ inch thick, were taken from each rail. The tensile test-pieces were 15 inches long, with a reduced section ~ inch in diameter, and 5 inches between shoulders. They also had a groove just beyond each end of th~ reduced section for holding the micrometer arrangement used in determining the elastic limit. The torsion test-pieces were of the usual size, 4 inches long and 1 inch square, with a reduced section ~ inch in diam-
Mar., lSSl.J
Dudley--Steel Rails.
181
eter, and 1 inch between shonlcters, with a ~ inch fillet. For the shearing test one-half of the tensile test-pieces, after they were broken, was turned down to a rod -~ inch in diameter, and sheared off (single shear) in a shearing apparatus prepared for the purpose. For the bending test two pieces 1½ inches wide, and 12 inches long, were slotted (~ut of the web of each rail. Accompanying Plate 1 gives the details of the form and size of the test-pieces and of the shearing apparatus. A and B represent the shearing apparatus, full size ; C represents the tensile test-piece, half size; D, the torsion piece, full size; E, the shearing piece, full size ; and F, the bending test-piece, half size. The half-inch section of each rail was used in determining the present weight per yard of the worn rails, the original Weight per yard of these rails when lald, as will be described further on, and also in making the diagrams of the worn rails, which appear in the accompanying Plates 2, 3, 4 and 5. When the tensile test-pleces had been prepared, one-half of them were annealed by heating them to a fair red heat, and then allowing them to cool slowly for 36 hours. Both the annealed and the unanhealed test-pieces were then tested in the tensile testing-machine. In determining the elastic limit yokes were fitted to the grooves in the test-pieces prepared for them, which yokes carried micrometer screws on opposite sides of the pieces, reading to the ten thousandth of an inch, and fitted up for electrical contact. The pieces were then strained with successive loads, increasing by 2000 pounds per square inch. After the application of each load the elongation was measured by the micrometers, and the point at which the elongation ceased to be directly proportioned to the load was regarded as the elastic limit. The results of these tests are given on Plates 6 and 7. Tim figures as to tensile strength and elastic limit mean pounds per square inch. The elongation is per cent. in five inches. Neither of the torsion testpieces were mmealed, they being tested in the condition in which the steel was when removed from the track. The diagrams made by the test of these pieces were measured up, and the results are given in connection with the results of the tensile tests as above described. The figures given as to length of diagram, height of diagram, and elastic limit are in inches and hundredths of an inch ; and those giving areas ~f diagram are square inches. They are the mean of the results from the two test-pieces. In the shearing tests the figures given under "shearing stress" are pounds per square inch ; ~hile the "detrusion "
182
Dudley--Steel Rails.
[5our. Frank.Inst.,
before rupture represents the travel or motion of the shearing piece A, in Plate 1, measured in decimals of an inch. In making these measurements a load of 2000 pounds was put oll the pieces and then a reading made ; at the moment of rupture another reading was made. The ,difference of these two readings represents the detrusion. In making the bending tests, the pieces before described were supported on knife edges 10 inches apart, and bent by a third knife edge at equal distances fl'om the supports, after the manner of making transverse tests of metal. In making these tests it was found that the load applied gradually increased until it reached a maxinmm, and then diminished slowly. The deflectiou obtained at the point of maximum load varied considerably in the different rails, being a half inch or more eacl(side of two inches. After the maximum load had beei~ obtained as above described, the pieces were removed from the strain, and set up on end in the same machine, and then bent until rupture took place or until they could be bent down no farther. The broken or bent pieces were then laid on a piece of paper in the positiou in which they were at rupture or as bent, and the amount of deflection from a straight line of the piece broken or bent was measured with a protractor. The figures given under the head "bending tests" are maximum load and deflection. The maximum load was calculated from the data obtained in testing the pieces, for a piece 1½ inches wide and ½ inch thick. The deflection is iu degrees. In all cases where the deflection is given as 190 °, the pieces did not break, but had the ibrm when removed from the machine of a letter U with the ends brought together until they formed with each other an angle of 10 °. From the torsion test-pieces after they were broken borings were taken for analysis. In these borings the carbon, phosphorus, silicon and manganese were determined. All the chemical work was done in duplicate. The carbon was determined, after separation from the iron by chloride of copper and ammonium, by combustion in oxygen gas, the phosphorus by the molybdate method, the manganese by the bromine method, and the silicon in the usual manner. The result of these determinations are given with the other data. So nmch for the methods of the chemical and physical testing of these rails. In determining the rate of wear of a steel rail it is of course necessary to know the loss of metal per yard which each rail has suffered and the tonnage. The loss of metal per yard divided by the tonnage
Mar., 1881.]
Dudley--Steel Rails.
183
gives the rate of wear ; and when this datum is obtained for a series of rails, it furnishes a means of comparison as to their wearing power. The tonnage of the 64 rails we are studying was computed from the data as to number of trains and movement of cars in the office of the superintendent of transportation, Mr. John Reilly. These data were twice worked over, and while I do not think that the tonnages given with each rail are absolutely the number of tons which have passed over them, I do think that the percentage error is small, and. that the comparison of one rail with another by means of these tonnages, which is really what we are after, gives results that call safely be relied on. With regard to the loss of metal per yard and the method by which. it was obtained it will be necessary to go a little into detail. Of course nothing could be simpler than to obtain the loss of metal per yard off these rails, provided the data were at hand, for this is simply the dif-ference between the present weight per yard of the worn rails and the original weight per yard of these rails when they were laid. But, unfortunately, these rails were not weighed when they were put in track~ and so one essential element of our dam is wanting. Nor will it do to assume that these rails originally weighed 67 lbs. or 56 lbs. per yard, which is the standard weight of the different rails embraced: in this series. ~[r. J. W. Cloud, engineer of tests, finds by the weight of a number of new rails just from the mills, weighed at Altoona during two or three years past, that they vary from } lb. to 1½ lbs. per yard from standard weight. Still further, an examination of the rails in our series shows that some of them~ from the web being thinner than standard, which means that the rolls were closer tbgether when that rail was rolled, could not have weighed when new more than 64 lbs.'or 65 lbs. per yard. In view of these statements, the question fairly meets us: How can the loss of metal per yard of these rails be determined? I t is evident to all, I think, that if we know the weight per yard of these worn rails, and then are able to obtain the areas (1) of a section of the worn rail, and (2) of a section of the original rall as rolled, we have at hand all the data necessary tbr obtaining the weight per yard of the rall as rolled. For it is clear, I am sure, that as the area of a section of the worn rail is to its weight per yard so is the area of a section of the original rail to its weight per yard. If, thereibre, we can obtain (1) the weight per yard of the worn rails, (2) the area of ~t
184
Dudley--Steel Rails.
[Jour. Frank. Inst.,
section of each of these rails, and (3) the area of a section of each rail as rolled, we shall be able to obtain the lo~ of metal per yard of each of the rails in the series we are studying. Can these data then be obtained ? First, as to the areas. How can the area of a section of the worn rail and of a section of the rail as rolled be obtained ? The following was the method used.. The half-inch section of the worn rails before referred to was laid upon a sheet of paper, and its outline traced upon the paper with the utmost care by means of a very sharp-pointed hard pencil. The surface inclosed within this tracing corresponds very closely with the section of the worn rail. Then directly over this tracing was laid the template of the section after which this rail was rolled, and a tracing then made with a sharp-pointed hard pencil as hefore of that portion of the head which was lacking in the wbrn rail. I f now the original height of the rail was the same as the template used, and if the original shape of the head of the rail was that of the template, we have in this manner a diagram on paper representing a section of the worn rail, and also one representing a section of the original rail as rolled. The diagrams obtained in this manner for each rail in the series are represented in accompanying Plates 2, 3, 4 and 5. The areas of these diagrams were then obtained by means of the planimeter. What now are the assumptions, and what are the probable errors in this method ? The first a~nmption is that the original rails as rolled were the same height as the template. This is probably not exactly true in flint. The wear of the rolls, and possibly carelessness in making the rolls originally, together with the ihct that the rails rolled in the same set of rolls were probably not all rolled at the same temperature, and consequently shrunk different amounts in cooling, might'each occasion small deviations from the height which the rail should have according to template. And yet Mr. Cloud, engineer of tests, has callipered the height of some 60 new rails here at Altoona during the year past, which rails were from different mills, and has found the variation from standard height not more than x ~ of an inch. It would seem, therefore, that the error arising from the assumption as to height of rail could not be large. The second assumption is that the original shape of the top of the head of the rails we are studying was that of the template. This is of course a question of accuracy in the original nmlmfaeture of the
5far., 1881.']
Dudley--Steel Rails.
185
rolls, and of their wear. :Now it is impossible to say how accurately the rolls were originally made, and how much they have worn out of shape when any raii was rolled. But in ally case the error must be very small, for any noteworthy deviation of the top of the head from smlldard wo'hld have given difficulty in securing a straight, even track, and would have caused t h e rejection of those rails by the inspector at the mills. The third assumption is that these rails have not suffered distortion as Co height by the service which they have endured. Upon this point I will say that I think an inspection of the diagrams in accompanying Plates 2, 3, 4 and 5 will convince any one that this distortion, if any, is excessively small. I t should be stated here that any deviation of ~hese rails, from template, such as beading over on one side of the head, or thicker or thinner webs than standard, or variation in the shape of the foot or in the under side of t h e head, can cause no error ; because in taking the areas with the planimeter, the diagrams from which the area of the worn section and the area of the original section were taken, coincide in every part except in that part of the original section which was traced-in fl'om template as above described. In other words, in getting the areas of the rails as rolled, the actual section of the rail just as it was rolled was used in every part, except that portion of the head which was golm from wear, which portion was restored from template. Fourth. H o w accurately can the areas be obtained by means of the planlmeter? The planimeter which was used gave readings to ]-~-~ of a square inch. If, therefore, the manipulation was such as to give the instrument its full chance, the maximum error in area could not be over , 10.0 of a square inch. As a matter of t~act, in taking tlle area~ with the planimeter, from three to five measuremeuts were made of each diagram, which measurement.s differed from each other not more than from one to three hundredths of an inch, and then a mean of these measurements was taken as representing the area of the diagram. I t seems probable, therefore, that the error arising from this cause could not in any case amount to more than one or two hundreths of a square inch. Now as to the weight per yard of the woru rails. How can this best be obtained? Obviously the most simple method would be to weigh each of the worn rails, and then divide the weight by the length. This was done with these rails, as has already been mentioned. But
186
Dudley--Steel Rails.
[Jour. Frank. Inst.,
when the results were obtained, and the original weight of the rails computed, by means of the areas as above described, it was found that some rails rolled at the same mill, at the same time, and with the same thickness of web and same shape of foot, differed from each other ill the original weight, as computed, from 1½ to 3 lbs. per~ard. The explanation of this seems to be that the half-inch sections before referred to, which were used in getting the areas, did not exactly represent the whole rail; in other words, the rail was unevenly worn. This explanation was confirmed by callipering the height of the halfinch sections, and comparing these heights with the heights of the rail in different places. I t became necessary, therefore, to devise some other means of obtaining the weight per yard of the worn rails, and it was finally decided to obtain the desired weights from the half-inch sections themselves before referred to. These sections were all, therefore, carefully weighed on a balance which weighed accurately to one-half a grain. If, now, the half-inch sections were exactly half an inch in thickness, a little consideration shows that it is possible in this way to get at the weight of a rail moreaccurately than it could be obtained by weighing the whole rail on a common platform scale, supposing, of course, that the rail was uniform in section throughout its length. For there are 72 half-inch sections in a yard, and in a 30-foot rail 720 half-inch sections. If, now, the error in weighing a half-inch section is one-half a grain, the total error in the weight of. a 30-foot rail would be 720 half grains, or 360 grains, which is a little less than an ounce. Inasmuch, therefore, as ordinary platform scales do not generally weigh closer than half a pound, it is evident that, so far as accuracy of weighing is concerned, the method of getting at the weight per yard of these worn rails, by weighing the half-inch sections, leaves nothing to be desired. But were these half-inch sections exactly half an inch thick ? An examination of these sections by means of vernier callipers, reading to one-thousandth of an inch, showed that they were not exactly half an inch thick, and that they varied in thickness in different parts of the section. The reasons for this seem to be, that it is almost impossible to set a tool so as to cut off" exactly half an inch from a rail, and that in cutting steel as hard as some of these rails were, the tool is apt to spring more or less, and thus give a section of varying thickness. It became necessary, therefore, to devise some method by means of' which the average thickness of these sections could be obtained. This was
.~[ar.,1881.]
Dudley--Steel Rails.
187
done as follows : The half-inch sections were all carefully weighed in distilled water, at a temperature of 66°F., on a balance weighing to half a grain. Tile difference between the weight so obtained and the weight of the same sections in ~tir gives the weight of a volume of water equal to the volume of the section. Now, this weight of water divided by the weight of a cubic inch of water, which was taken as 252"5 grains, gives the volume of space occupied by any section. This volmne being known, it is of course only necessary to divide the same by the area obtainc
188
Dudley--Steel Rails.
[Jour. Frank. Inst.,
nage, the chemical analyses, the weights per yard, and loss of metal, as well as the loss per million tons, and the complete results of the physical tests and the density, are given in tabular form in accompanying Plates 6 and 7. These tables will be discussed farther on. The density is the weight of a cubic inch of the steel. This weight was obtained by dividing the weight of each of the half-inch sections by its volume, the data for this purpose having been obtained as previously described. The figures given Under density are fractions of a pound. In looking over the original weights of the rails as obtained, it will doubtless be observed that, although these rails were all supposed to have weighed 67 lbs. or 56 lbs. per yard when rolled, the figures given differ both ways from these figures. These differences are principally to be accounted for by differences in the shape of the bases and differences in the thickness of the webs. With regard to the latter point, it is evident that of two rails rolled at the same mill, in dae same year, if one has a web three or five one-hundredths of an inch thicker than another--which means that the rolls were farther apart when the thick-web rail was rolledmthe thin-web rail will be lighter than the other. The thickness of the webs of all the rails in d~e series we are studying was carefully measured with vernier callipers, and the differences from the standard thickness were found to vary both ways, from nothing up to five one-hundreths of an inch thinner than standard, and seven one-hundredths thicker than standard. Finally, the differences in the density of the different rails still further helps us to account for the differences in weight per yard of the rails as rolled. The following is the history and tonnage of the different rails : No. 881. Steel of 1868. Was on tangent, north rail, north track, in Bennington Cut. In service from June, 1888, to July, 1879-11 yrs. 1 mo. Grade, 92"4 ft. to the mile. Tonnage, 55,546,811 tons,
No. 882. Steel of 1868. Was oil tangent, north rail, north track, in Bennington Cut. I n service from June, 1868, to July, 1879-11 yrs. 1 too. Grade, 92"4 ft. to the mile. Tonnage, 55,546,811 tons. No. 883. Steel of 1869. Was on tangent, north rail, north track, on Whip-poor-will Straight. In service from May, 1869, to July,
Mar., 1881.]
Dudley~Sleel Rails.
189
1879--10 yrs. 2 mos. Grade, 95"04 ft. to the mile. Tonnage, 52,174,969 tons. No. 884. Steel Of 1869. Was on tangent, north rail, north track, on Whip-poor-will Straight. In service from M:ay, 1869, to July, 1879--10 yrs. 2 mos. Grade, 95"04 ft. to the mile. Tonnage+ 52,174,969 tons. No. 885. Steel of 1868. Was on tangent, south rail, north track, just west of Allegrippus Station. In service from July, 1868, to July, 1879--11 yrs. Grade, 89"76 it. to the mile. Tonnage, 55,197,99~ tons. No. 886. Steel of 1868. Was on tangent, north rail, north track, just west of A!legrippus Station. In service from July, 1868, to July, 1879--11 yrs. Grade, 89"76 ft. to the mile. Tonnage, 55,197,994 tOlls. No. 887. Steel of 1868. Was on tangent, south rail, north track, west of Allegrippus Station. In service from July, 1868, to July,: 1879--11 years. Grade, 89'76 ft. to the mile. Tonnage, 55,197,994 tons. No. 888. Steel of 1868. Was on tangent, south rail, north track, west of Allegrippus Station. In service from July, 1868, to July, 1879--11 years. Grade, 89'76 ft. to the mile. Tonnage, 55,197,994 tons. No. 889. Steel of 1873. Was on tangent, south rail, south track, at South Fork. In service fl'om August, 1873, to July, 1 8 7 9 - 5 yrs. 11 mo~. Grade, 21'13 ft. to the mile. Tonnage 44,620,100 to~s. No. 890. Steel of 1873. Was on tangent, north rail, south track, at South Fork. In service from August, 1873, to July, 1879-5 yrs. 11 mos. Grade, 21"13 ft. to the mile. Tonnage, 44,620,100 tons. No. 891. Steel of 1872. Was oll tangent, south rail, south track, at Summer Hill water plug. In service from June, 1872, to July, 1879--7 yrs. 1 too. Gra.de, 40"13 ft. to the mile. Tonnage, 53,687,192 tons. No. 892. Steel of 1872. Was on tangent, south rail, south track, at Summer Hill water plug. In service from June, 1872, to July, 1879--7 yrs. 1 too. Grade, 40"13 ft. to the mile. Tonnage, 53,687,192 tons. No. 893. Steel of 1874. Was on tangent~ north rail, south track~
190
Dudley--Steel Rails.
[Jour. Frank. Inst.,
near Summer Hill. In service from May, 1874, to July, 1879-5 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tonnage, 38,088,572 tOlls. ~No. 894. Steel of 1874. Was on tangent, south rail, south track, east of Portage. In service from May, 1874, to July, 1879--5 yl~. 2 mos. Grade, 52"8 ft. to the mile. Tonnage, 38,088,572 tons. No. 895. Steel of 1874. Was on tangent, south rail, south track, east of Portage. In service from May, 1874, to July, 1879--5 yrs. 2 mos. Grade, 52"8 ft. to the mile. Tonnage, 38,088,572 tons. No. 896. Steel of 1874. Was on tangent, south rail, south track, east of Portage. Iu service from May, 1874, to July, 1879--5 yrs. 2 mos. Grade, 52"8 ft. to the mile. Tonnage, 38,088,572 tons. No. 897. Steel of 1868. Was on high side of 5 ° curve, north rail, north track, east of Bridge :No. 3, Summer Hill. In service from May, 1868, to July, 1879,--11 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tonnage, 52,370,617 tons. No. 898. Steel of 1868. Was on high side of 5 ° curve, north rail, north track, east of Bridge No: 3, Summer Hill. In service from May, 1868, to July, 1879--11 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tolmage, 52,370,617 tons. No. 899. Steel of 1868. Was on low side of 5 ° curve, south rail, north track~ east of bridge No. 3, Summer Hill. In service from May, 1868, to July, 1879--11 yrs. 2 mos. Grade, 21"12 i~. to the mile. Tommge, 52,370,617 tons. No. 900. Steel of 1868. Was on low side of 5 ° curve, north rail, north track, east of Bridge No. 3, Summer Hill. In service from May, 1868, to July, 1.879--11 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tonnage, 52,370,617 tons. ~'o. 901. Steel of 1871. Was on low side of 5 ° curve, south rail, north track, in Gable's Cut, near Summer Hill. In service from August, 1871, to July, 1879--7 yrs. 11 inos. Grade, 39'6 ft. to the mile. Tonnage, 40,061,230 tons. No. 902. Steel of 1871. Was on high side of 5 ° curve, north rail, north track, in Gable's Cut, near Summer Hill. In service from August, 1871, to July, 1879--7 yrs. 11 mos. Grade, 39"6 ft. to the mile. Tonnage, 40,061,230 tons. No. 903. Steel of 1876. Was on high side of 5 ° curve, south rail, south track, west of Bridge No. 1, Summer Hill. In service from
Mar., 1881.]
Dudley--Steel Rails.
191
August, 1876, to July, 1879--2 yrs. 11 mos. Grade, 39"6 ft. to the mile. Tonnage, 21,504,824 tons. No. 904. Steel :of 1876. Was on low side of 5 ° curve, north rail, south track, west of bridge No. 1, near Summer Hill. In service from August, 1876, to July, 1879--2 yrs. 11 mos. Grade, 39"6 ft. to the mile. Tonnage, 21,504,824 tons. No. 905. Steel of 1868. Was on high side of 5 ° curve, north rail, north track, east end of Milmore Middle Siding. Ill service from June, 1868, to April, 1879--10 yrs. 10 mos. Grade, 47'52 ft. to the mile. Tonnage, 50,648,939 tons. ~o. 906. Steel of 1868. Was on low side of 5 ° curve, south rail, north track, east end of Wilmore Middle Siding. In service from June, 1868, to April, 1879--10 yrs. 10 mos. Grade, 47"52 ft. to the mile. Tonnage, 50,64S,939 tons. No. 907. Steel of 1874. Was on high side of 5 ° curve, north rail, north track, east of Wilmore Middle Siding. In service from June, 1874, to April, 1879--4 yrs. 10 mos. Grade, 47"52 ft. to the mile. Tonnage, 24,211,147 tons. No. 908. Steel of 1875. Was on low side of 4 ° curve, north rail, south track, near tower at east end of Ga]itzin Tunnel. In service . fi'om May, 1875, to July, 1879--4 yrs. 2 mos. Grade, 95"04 feet to the mile. Tonnage, 32,428,6]4 tons. No. 909. Steel o f 1875. Was on Mgh side of 4 ° curve, south rail, south track, near tower at east end of Galitzin Tunnel. In service from May, 1875, to July, 1879---4 yrs. 2 mos. Grade, 95"04 ft. to the mile. Tonnage, 32,428,614 tons. No. 910. Steel of 1871. Was on low side of 5 ° curve, south rail, south track, west of Allegrippus Station. In service from July, 1871, to July, 1879--8 yrs. Grade, 89"76 ft. to the mile. Tonnage, 62,813,664 tons. No. 911. Steel of 1877. Was on low side of 5 ° curve, south rail, south track, second curve west of Allegrippus Tower. ]n service from April, 1877, to July, 1879--2 yrs. 3 mos. Grade, 89"76 i't. to the mile. Tolmage, 17,226,993 tons. No. 912. Steel of 1873. Was on high side of 5 ° curwi north rail, south track, second curve west of Allegrippus Tower. In service from July, 1873, to July, 1879--6 yrs. Grade , 89"76 ft. to the mile. Tonnage, 47,438,145 tons. No. 913. Steel of 1870. Was on tangent, north rail, north track,
192
Dudley--Steel Rails,
[Jour. Frank. Inst.,
600 feet west of mile post 211 from Philadelphia. In service from March, 1870, to July, 1879--9 yrs. 4 mos. On level. Tonnage, 45,855,101 tons. No. 914. Steel of -1870. Was on tangent, south rail, north track, 600 feet west of mile post 211 from Philadelphia. In service from March, 1870, to July, 1879--9 yrs. 4 mos. On level. Tonnage, 45,855,101 tons. No. 915. Steel of 1873. Was o n tangent, south rail, north track, east of Petersburg Toolhouse. In service from June, 1873, to July, 1879--6 yrs. 1 too. On level. Tonnage, 31,514,889 tons. No. 916. Steel of 1873. Was on tangent, north rail, north track, just west of Ardenheim. In Service from July, 1873, to July, 1879 - - 6 yrs. On level. Tonnage, 31,127,829 tons. No. 917. Steel of 1873. Was on tangent, south rail, north track, just west of Ardenheim. In service from July, 1873, to July, 1879 --6 yrs. On level. Tonnage, 31,127,829 tons. No. 918. Steel of 1869. Was on tangent, north rail, north track, west of Vandevaner's Bridge. I l l service from January, 1869, to July, 1879--10 yrs. 6 mos. On level. Tonnage, 51,720,011 tons. No. 919. Steel of 1869. Was on tangent, north rail, north track, west of Vandevaner's Bridge. In service from January, 1869, to July, 1879--10 yrs: 6 mos. On level. Tonnage, 51,720,011 tons. No. 920. Steel of 1867. Was on tangent, south rail, north track, at Jackstown Water-trough. In service fi'om December, 1867, to July, 1879--11 yrs. 7 mos. On level. Tonnage, 52,991,684 tons. No. 921. Steel of ]874. Was on tangent, north rail, north track, 200 feet east of mile post 211 from Philadelphia. In service from March, 1874, to July, 1879--5 yrs. 4 mos. On level. Tonnage, 27,622,230 tons. .No. 922. Steel of 1874. Was on tangent, south rail, north track, 200 ft. east of mile post 211 from Philadelphia. In service from March, 1874, to July, 1879--5 yrs. 4 mos. On level. Tonnage, 27,622,230 tons. No. 923. Steel of 1873. Was on tangent, south rail, north track, east of Petersburg Toolhouse. In service from June, 1873, to July, 1879--6 yrs. 1 mo. On level. Tonnage, 31,514,889 tons. No. 924. Steel of 1875. Was on taugent, south rail, south track, west of Vandevaner's Bridge. Iu service from June, 1875, to July, 1879--4 yrs. 1 mo. On level. Tounage, 36,349,989 tons.
Mar., 1881.]
Dudley--Steel Rait~.
193
No. 925. Steel of 1875. Was on tangent, south rail, south track, west of Vandevaner's Bridge. 1,1 service from June, 1875, to July, 1879--4 yrs. 1 too. On level. Tonnage, 36,349,989 tons. No. 926. Steel of 1870. Was on tangent, south rail, south track, 1500 ~eet west of mile post 182 from Philadelphia. In service from April, 1870, to July, 1879--9 yrs. 3 mos. On level. Tonnage, 76,409,123 tons. No. 927. Steel of 1870. Was on tangent, south rail, south track, 1500 feet west of mile post 182 from Philadell~fia. In service from April, 1870, to July, 1879--9 yrs. 3 mos. On level. Tonnage, 76,409,123 tons. No. 928. Steel of 1874. Was on tangent, south rail, south track, 200 feet east of mile post 181 from Philadelphia. In service fl'om JuDy, 1874, to July, 1879--5 yrs. On level. Tonnage, 43,610,150 tons. No. 929. Steel of 1867. Was on high side of 2¼° curve, north rail~ north track, 250 feet east of mile post 194 from Philadelphia. In service from December: 1867, to July, 1879--11 yrs. 7 mos. On level. Tonnage, 52,991,684 tons. No. 9:30. Steel of 1867. Was on low side of 2} ° curve, south rail, north track, 250 feet east of mile post 194 from Philadelphia. In service from December, 1867, to July, 1879--11 yrs. 7 mos. On level. Tonnage, 52,991,684 tons. No. 931. Steel of" 1868. Was on high side of 3 ° curve, south rail, north track, 400 feet east of mile post 191 from Philadelphia. In service from January, 1869, to July, 1879--10 yrs. 6 mos. On level. Tonnage, 51,720,011 tons. No. 932. Steel of 1868. Was on low side of 3 ° curve, north rail, north track, 400 feet east of mile post 191 from Philadelphia. In service from January, 1869, to July, 1879--10 yrs. 6 mos. On level. Tonnage, 51,720,011 tons. No. 933. Steel of 1869. Was on high side of 3} ° curve, north rail, south track, 1000 feet east of mile post 183 from Philadelphia. In service t~om November, 1869, to July, 1879--9 yrs. 8 mos. On level. Tonnage, 78,364,968 tons. No. 934. Steel of 1869. Was on low side of 3} ° curve, south rail, south track, 1000 i~ct east of mile post 183 from Philadelphia. In ~rvice from November, 1869, to July, 1879--9 yrs. 8 mos. On level. Tounagc, 78,364,968 tons. WHOLE:No.VoL. CXI.--(THIR]) SERI~S.Vol. lxxxi.) 13
194
Dudley--JS'ieel Rails.
[Jour. Frank.Inst.,
:No. 935. Steel of 1866. Was on high side of 2 ° curve, north rail, south track, at McVeytown Station. In service from September, 1866, to June, 1879~12 yrs. 9 mos. On level. Tonnage, 92,025,478 tons. :No. 936. Steel of 1866. Was on low side of 2 ° curve, soutll rail, south track, at McVeytown Station. In service from September, 1866, to June, 1879--12 yrs. 9 mos. On level. Tonnage, 92,025,478 tt)n No. 937. Steel of 1~876. Was on low side of 2 ° curve, south rail, south track, 200 feet west of mile post 192 from Philadelphia. In service from March, 1876, to July, 1879--3 yrs. 4 mos. On level. Tonnage, 29,905,122 tons. No. 938. Steel of 1876. Was on high side of 2 ° curve, north rail, south track, 200 feet west of mile post 192 from Philadelphia. In service from March, 1876, to July, 1879--3 yrs. 4 mos. On level. Tonnage, 29,905,122 tons. :No. 939. Steel of 1875. Was on the low side of 4 ° curve, south rail, south track, 250 feet east of Mount Union Bridge. In service from May, 1875, to July, 1879--4 yrs. 2 mos. On level. Tonnage, 37,150,179 tons. No. 940. Steel of 1875. Was on the high side of 4 ° curve, north rail, south track, 250 feet cast of Mount Ullion Bridge. In service from May, 1s75, to July, 187.9--4 yrs. 2 mos. On level. Tonnage, 37,150,179 tons. No. 941. Steel of 1874. Was on high side of 3 ° curve, south rail, south track, 300 feet east of Manayunk. In service from September, 1874 to July, 1879--4 yrs. 10 mos. On level. Tonnage, 42,277,638 tons. No. 942. Steel of 1874. Was on low side of 3 ° curve, north rail, south track, 300 feet east of Manayunk. In service from September, 1874, to July, 1 8 7 9 - 4 yrs. 10 mos. On level. Tonnage, 42,277,638 tons. No. 943. Steel of 1873. Was on the low side of 2 ° curve, south rail, south track, 2000 feet west of mile post 179 t~'om Philadelphia. Ill service from April, 1873, to July, 1879--6 yrs. 3 mos. On level. Tonnage, 55,127,464 tolls. No. 944. Steel of 1873. Was oil high side of 2 ° curve, north rail. south track, 2000 feet west of mile "post 179 from Philadelphia. In
Mar., 1881.]
.Dudley--Steel Rails.
195
service from April, 1873, to July, 1879--6 yrs. 3 mos. On level. Tonnage, 55,127,464 tons. Accompanying Plates 6 and 7 give in tal)ular form all the essential data of the history of the rails, together with the chemical analyses, weights, and results of the chemical tests, as has been before stated. To a study of these tables attention is now directed. As will be observed, the rails have been arranged in six groups, according to the kind of service to which they have been subjected. The first group of 16 rails did their service on grade tangents; the second group of 8 rails on the low side of grade curves; the third group of 8 rails on the high side of grade curves; the fourth group of 16 rails on level tangents ; the fifth group of 8 rails on the low side of level curves ; and the sixth group of 8 rails on the high side of level curves. As will likewise be observed, the rails within the groups have been arranged in regular order according to loss of' metal per million tons, the rail having lost least metal per million tolls being placed first. The first rails in the several groups are therefore the slower wearing, while the latter rails are the more rapid wearing. This arrangement enables us to compare the slower-wearing rails of the different kinds of service with the more rapid-wearing ones, and to discover what physical qualities, and what chemical composition, are characteristic of the slower-wearing as well as the faster-wearing rails. Giving our attention now to the tables, I think the first observation will be that there is no absolute gradation in physical qualities, or in chenfical composition, applying to every rail in each group, which corresponds to the gradation in amount of metal lost per million tolls. In other words, if we take tensile strength, or elongation in the physical qualities, or the carbon or phosphorus in the chemical composition, we do not find that the first rail in each group is characterized by a certain figure in any one or nmre of these respects, while the last rail in the group is characterized by another different figure, and all the intermediate rails arrange themselves in respect to any of these peculiarities, between these two extreme rails of the group. Nor do I think we ought to expect such uniibrmity, aud for the ibllowing reasons : First. Whatever errors there may have been in determining the loss of metal of these rails or in the tonnages, as has been before described, will of course have an influence in determining the position
196
Dudley--Steel Rails.
[Jour. Frank. Inst.~
of any rail in the group to which it belongs. And, as has already been said, while I do not think that these errors, whatever they may be, are sufficient to seriously obscure or counteract the general conclusions which the results are calculated to teach, yet these errors may possibly be large enough to give some rail a position in the groul~ which it should not occupy. Second. I am not aware that it is known as yet exactly what wear is, or what it is dependent upon. Some considerations in regard t~ the nature of wear will be advanced further on ; but whether wear is a direct function of the tensile strength of steel, or of its elongation~ or of its elastic limit, or of its resilience, or any combination of these, or indeed, as seems somewhat probable, of the amount of distortion by bending that a piece of steel will suffer, is a problem yet to be solved. I t may be that wear bears a direct ratio to the elongation within the elastic linfit, or to the amount of work done within the elastic limit whenever a particle is worn off; or, indeed, wear may be a direct function of the granular structure of steel, the wear being more rapid as the granular structure is coarser, or vice versa. Or, finally, wear may be due to a combinations of causes, and each cause may require for its elucidation a different physical test. I t is perhaps, therefore, not strange, in view of' this want of knowledge, that, in a series of steels arranged according to h)ss of metal, or on the principle of wear, the physical qualities which we are now able to measure should not show uniform gradations throughout the series. But, it seems t~ me, this reasoning does not tend to throw doubt on our ability to draw conclusions that will be valuable from the work which we have in hand. :For the question we are studying is not what does wear depend upon, but what chemical composition, and what physical properties, are, in general, characteristic of such rails as have in actual service given least loss of metal per million tons' burden ? While the answer to this question may not solve the whole problem of wear, yet it cannot fail to throw light on the relation between the chemistry and physics of steel and its wearing power, which, after all, with our present knowledge of steel metallurgy, is all the knowledge we are able to utilize. (To be continued.)
Dudley--Steel Rails.
177
tile combustible portion Of "Nixon's navigation coal" was intrinsically /12"7519--12"0906 X 1 0 0 = ~ 5"47 per centum superior in water \ ! 12"0906 vaporizative power to the pound of the combustible portion of the Cumberland semi-bituminous coal. That is to say, if both coals had the same percentage and kind of refuse such would be their relative value in, water heating effect. This difference may very easily be allowed. The very great difference in the water vaporizative power of equal weights of the two crude coals prevents their use for measuring the economic performances of themachinery in the two cases.
THE W E A R I N G P O W E R O F S T E E L R A I L S I N R E L A T I O N TO T H E I R C H E M I C A L C O M P O S I T I O N A N D PHYSICAL PROPERTIES.* By CHARLES B. D U D L E Y , Ph.D., Chemist Pennsylvania Railroad Company, Altoona, Pa. Read at the Philadelphia Meeting of the American Institute of Mining Engineers, held at the Franklin Institute, February 17, 1881.
TREO. N. ELY, Esq., Superintendent of Motive Power, Pennsylvania Railroad. DEAR S,R--It is now nearly three years since my first report to you on the subject of steel rails was written. That report, as you will At the Lake George meeting of the Institute in October, 1878, I had the honor of presenting to the [nstitute, through the kind permission of the officers of the Pennsylvania Railroad Company, the results of a study of twenty-five examples of steel rails, which had all been in actual service. Considerable discussion followed tile publieation of that paper, and there seemed to be a strong disinclination, especially on the part of the steel rail manufacturers, to accept the conclusions presented. One of the principal objections urged against the conclusions drawn was that they were based on too few samples; in other words, that no conclusions safe to act upon could be drawn from the examination of twenty-five rails. In view of this criticism it wm~ decided to repeat the investigation with a larger number of samples, and with the aid of the experience gained in the first investigation. The results of this Second study of steel rails are, by the permission of the officers of the Pennsylvania :Railroad Company, herewith presented to the Institute, with the sincere desire that they may aid in adding to our knowledge of this most important product. Like the previous paper, this is in the form of a report to one of the officers of the company, which will account lbr the style, and the manner of presenting tile data. WHOLE ~O. VOL. CXI.--(TtHRD SERIES, Vol. lxxxi.) 12
178
Dudley--Steel Rails.
[Jour. Frank. Inst,
remember, dealt principally with the question of the relation between the chemical composition and physical properties of steel rails and :their power to resist crushing and fracture in actual service. Other matters were referred to or touched upon in that report, but the main question was, Why do some rails crush or break in service while ,others do not ? You will doubtless remember that the principal eon,clusion arrived at was, that the softer rails are less liable to crush or break in service than the hard ones. Or, in other words, so far as ~onclusions could be drawn fi'oln the chemical analysis and physical ~est of 25 samples of steel rails which had actually been in service, these conclusions were that those rails which have the smaller amounts of carbon, plmsphorus, silicon and manganese are less liable to crush ,or break in service than those which have larger amounts of these ,elements. Or, again, looked at in the light of physical tests, those rails which have the lower tensile strength and the greater elongation are the ones which give the least trouble from breaking or crushing in ~rack. In the report just referre~l to, the question of the wearing power of steel rails was not made prominent, and from that report no positive :and definite information could be obtained as to what quality of steel would give rails that would endure the greatest amount of traffic with £he least loss of metal. And yet, of the three principal causes which ~ccasion the removal of rails from the track, viz., broken, crushed and worn out, perhaps the latter is of the most importance'. With ~he improvement in maintenance of way which has characterized the Pennsylvania Railroad during the last five or six years, the removal .of rails from track from the first two of these causes has, if I am right, quite notably diluinshed. This certainly is true with regard to broken rails. And if, as time advances, the number of crushed rails ,shall dhninlsh, both because of the continued improvement in main~euance of way, before referred to, and because, owing to improved and better methods at the steel works, there are fewer crushed rails caused by physical defects in the steel, the question of the wearing power of steel rails ohviously becomes the all-important one. In view of these considerations, it was thought that an investigation into the relation between the wearing power of steel rails and their chemical composition and physical properties could not fail to throw light upon a question of vital importance in the nmnagement of the Pennsylvania Railroad. - The results of such an investigation are presented in the
Mar., 1881.]
Dudley--Steel Rails.
179
ibllowing report, which deals not, as did the previous report, with the relations between the chemical and physical characteristics of steel rails and their power to resist crushing or fracture, but entirely with the relation between these characteristics and wear. On ally railroad the rails are or may be called upon to perform their service under quite varying conditions. Oil a railroad like the Pennsylvani'a Railroad, for example, there are levels and grades, and there are tangents and curves, and tllcre are combinations of the levels and grades with the tangents and curves. Moreover, the rails oll the high sides of curves do their service under different conditions from those on the low sides of curves. So that, as far as kind of service is concerne(i, a rail may be called upon to do its work under one of these six conditions, viz., on level tangent, on high side of level curve, oil low side of level curve, on grade tangent, on high side of grade curve, or on low side of grade curve. Any investigation, therefore,into the wearing power of steel rails, which shall be of service in determining what rails are best for the whole road, must take into account these slx conditions. O f course there arc many different grades and curves of differeut radius on the Pennsylvania Railroad ; but to study wear on each grade, and for every radius of curvature, would make a problem almost liih-long. In order, therefore, to make the work manageable it becomes necessary to choose some average grade and some medium degree of curvature, as representing the dittbrent grades and curves on the road. This has been done in the work about to be described. Now, the problem before us is: What chemical composition and w~mt physical properties are characteristic of those rafts which in actual service have lost least metal in i,roportiou to the tonnage that has passed over them? In order to get the information neces~ry to answer this question, 64 rails were taken from the track in July, 1879, and subjected to chenfical analysls and physical test, as is detailed ihrthe!" Oil. Sixteen of these rails were taken from level tangents and 16 from level curves, 8 from the high side and 8 from the low side of the curves. Again, 16 rails were taken from grade tangents and 16 from grade curves, 8 from the high side arid 8 from the low side of these cm'ves. The 32 rails on grades were all taken between Conemaugh and Altooua, and the 32 on levels beCween Tyrone and Miftlin. Furthermore, 32 of the rails were taken fl'om the north track, and 32 from the south track. The principle governing the selection was to
180
Dudley--Steel Rails.
[Jour. Frank. Inst..
secure 8 as slow-wearing rails and 8 as rapid-wearing rails as could be found on each of the fimr conditions of service--level tangents, level curves, grade tangents and grade curves--the rails on curves bein~ taken, as has been already described, one-half from the high side or' the curve and one-half from the low side of the curve. In the actual selection a pair of callipers was used, and the loss in height of a rail, compared with it~ time of service, gave a su~eiently accurate measure of the rate of wear of a rail to determine its selection. It should be stated here that since the rails were selected in July, after the usual annual removal from the track of worn-out rails, which takes place earlier in the season, the more rapid-wearing rails obtained for examination are not as marked examples of rapid-wearlng rails as would have been obtained had the selection been earlier made. With regard to the 64 rails chosen, it may also be stated that not one of them had broken in service, and only one showed any signs of crushing ; so that, as far as every quality is concerned, except their rate of wear, ali of them might be classed as good rails. It was thought that a chemica[examination and physical test of the 64 rails, compared with the loss of metal they had suffered by the tonnage which had pa;sed over them, could not Sail to throw some light on the qnestion as to what kind of steel in rails gives best wear. These rails having been selected were removed from the track and sent to Altoona. They were then cleaned and weighed on an ordinary platform scale, their length measured as accurately as possible with a steel tape, and their height callipered near both ends and in the middie. The weight of the whole rail divided by its length obviously gives the present weight per yard of these worn rails. The weights obtainltl in this manner, however, were not subsequently used, as will be explained further on. After the weighing, five feet were cut off from the end of each rail for test purposes. The test-pieces, except those used in the bending test, were all cut from the head of the rail. Two pieces for tensile test, two for torsion test, and two pieces from the web for bending test, as well as a section of the rail ½ inch thick, were taken from each rail. The tensile test-pieces were 15 inches long, with a reduced section ~ inch in diameter, and 5 inches between shoulders. They also had a groove just beyond each end of th~ reduced section for holding the micrometer arrangement used in determining the elastic limit. The torsion test-pieces were of the usual size, 4 inches long and 1 inch square, with a reduced section ~ inch in diam-
Mar., lSSl.J
Dudley--Steel Rails.
181
eter, and 1 inch between shonlcters, with a ~ inch fillet. For the shearing test one-half of the tensile test-pieces, after they were broken, was turned down to a rod -~ inch in diameter, and sheared off (single shear) in a shearing apparatus prepared for the purpose. For the bending test two pieces 1½ inches wide, and 12 inches long, were slotted (~ut of the web of each rail. Accompanying Plate 1 gives the details of the form and size of the test-pieces and of the shearing apparatus. A and B represent the shearing apparatus, full size ; C represents the tensile test-piece, half size; D, the torsion piece, full size; E, the shearing piece, full size ; and F, the bending test-piece, half size. The half-inch section of each rail was used in determining the present weight per yard of the worn rails, the original Weight per yard of these rails when lald, as will be described further on, and also in making the diagrams of the worn rails, which appear in the accompanying Plates 2, 3, 4 and 5. When the tensile test-pleces had been prepared, one-half of them were annealed by heating them to a fair red heat, and then allowing them to cool slowly for 36 hours. Both the annealed and the unanhealed test-pieces were then tested in the tensile testing-machine. In determining the elastic limit yokes were fitted to the grooves in the test-pieces prepared for them, which yokes carried micrometer screws on opposite sides of the pieces, reading to the ten thousandth of an inch, and fitted up for electrical contact. The pieces were then strained with successive loads, increasing by 2000 pounds per square inch. After the application of each load the elongation was measured by the micrometers, and the point at which the elongation ceased to be directly proportioned to the load was regarded as the elastic limit. The results of these tests are given on Plates 6 and 7. Tim figures as to tensile strength and elastic limit mean pounds per square inch. The elongation is per cent. in five inches. Neither of the torsion testpieces were mmealed, they being tested in the condition in which the steel was when removed from the track. The diagrams made by the test of these pieces were measured up, and the results are given in connection with the results of the tensile tests as above described. The figures given as to length of diagram, height of diagram, and elastic limit are in inches and hundredths of an inch ; and those giving areas ~f diagram are square inches. They are the mean of the results from the two test-pieces. In the shearing tests the figures given under "shearing stress" are pounds per square inch ; ~hile the "detrusion "
182
Dudley--Steel Rails.
[5our. Frank.Inst.,
before rupture represents the travel or motion of the shearing piece A, in Plate 1, measured in decimals of an inch. In making these measurements a load of 2000 pounds was put oll the pieces and then a reading made ; at the moment of rupture another reading was made. The ,difference of these two readings represents the detrusion. In making the bending tests, the pieces before described were supported on knife edges 10 inches apart, and bent by a third knife edge at equal distances fl'om the supports, after the manner of making transverse tests of metal. In making these tests it was found that the load applied gradually increased until it reached a maxinmm, and then diminished slowly. The deflectiou obtained at the point of maximum load varied considerably in the different rails, being a half inch or more eacl(side of two inches. After the maximum load had beei~ obtained as above described, the pieces were removed from the strain, and set up on end in the same machine, and then bent until rupture took place or until they could be bent down no farther. The broken or bent pieces were then laid on a piece of paper in the positiou in which they were at rupture or as bent, and the amount of deflection from a straight line of the piece broken or bent was measured with a protractor. The figures given under the head "bending tests" are maximum load and deflection. The maximum load was calculated from the data obtained in testing the pieces, for a piece 1½ inches wide and ½ inch thick. The deflection is iu degrees. In all cases where the deflection is given as 190 °, the pieces did not break, but had the ibrm when removed from the machine of a letter U with the ends brought together until they formed with each other an angle of 10 °. From the torsion test-pieces after they were broken borings were taken for analysis. In these borings the carbon, phosphorus, silicon and manganese were determined. All the chemical work was done in duplicate. The carbon was determined, after separation from the iron by chloride of copper and ammonium, by combustion in oxygen gas, the phosphorus by the molybdate method, the manganese by the bromine method, and the silicon in the usual manner. The result of these determinations are given with the other data. So nmch for the methods of the chemical and physical testing of these rails. In determining the rate of wear of a steel rail it is of course necessary to know the loss of metal per yard which each rail has suffered and the tonnage. The loss of metal per yard divided by the tonnage
Mar., 1881.]
Dudley--Steel Rails.
183
gives the rate of wear ; and when this datum is obtained for a series of rails, it furnishes a means of comparison as to their wearing power. The tonnage of the 64 rails we are studying was computed from the data as to number of trains and movement of cars in the office of the superintendent of transportation, Mr. John Reilly. These data were twice worked over, and while I do not think that the tonnages given with each rail are absolutely the number of tons which have passed over them, I do think that the percentage error is small, and. that the comparison of one rail with another by means of these tonnages, which is really what we are after, gives results that call safely be relied on. With regard to the loss of metal per yard and the method by which. it was obtained it will be necessary to go a little into detail. Of course nothing could be simpler than to obtain the loss of metal per yard off these rails, provided the data were at hand, for this is simply the dif-ference between the present weight per yard of the worn rails and the original weight per yard of these rails when they were laid. But, unfortunately, these rails were not weighed when they were put in track~ and so one essential element of our dam is wanting. Nor will it do to assume that these rails originally weighed 67 lbs. or 56 lbs. per yard, which is the standard weight of the different rails embraced: in this series. ~[r. J. W. Cloud, engineer of tests, finds by the weight of a number of new rails just from the mills, weighed at Altoona during two or three years past, that they vary from } lb. to 1½ lbs. per yard from standard weight. Still further, an examination of the rails in our series shows that some of them~ from the web being thinner than standard, which means that the rolls were closer tbgether when that rail was rolled, could not have weighed when new more than 64 lbs.'or 65 lbs. per yard. In view of these statements, the question fairly meets us: How can the loss of metal per yard of these rails be determined? I t is evident to all, I think, that if we know the weight per yard of these worn rails, and then are able to obtain the areas (1) of a section of the worn rail, and (2) of a section of the original rall as rolled, we have at hand all the data necessary tbr obtaining the weight per yard of the rall as rolled. For it is clear, I am sure, that as the area of a section of the worn rail is to its weight per yard so is the area of a section of the original rail to its weight per yard. If, thereibre, we can obtain (1) the weight per yard of the worn rails, (2) the area of ~t
184
Dudley--Steel Rails.
[Jour. Frank. Inst.,
section of each of these rails, and (3) the area of a section of each rail as rolled, we shall be able to obtain the lo~ of metal per yard of each of the rails in the series we are studying. Can these data then be obtained ? First, as to the areas. How can the area of a section of the worn rail and of a section of the rail as rolled be obtained ? The following was the method used.. The half-inch section of the worn rails before referred to was laid upon a sheet of paper, and its outline traced upon the paper with the utmost care by means of a very sharp-pointed hard pencil. The surface inclosed within this tracing corresponds very closely with the section of the worn rail. Then directly over this tracing was laid the template of the section after which this rail was rolled, and a tracing then made with a sharp-pointed hard pencil as hefore of that portion of the head which was lacking in the wbrn rail. I f now the original height of the rail was the same as the template used, and if the original shape of the head of the rail was that of the template, we have in this manner a diagram on paper representing a section of the worn rail, and also one representing a section of the original rail as rolled. The diagrams obtained in this manner for each rail in the series are represented in accompanying Plates 2, 3, 4 and 5. The areas of these diagrams were then obtained by means of the planimeter. What now are the assumptions, and what are the probable errors in this method ? The first a~nmption is that the original rails as rolled were the same height as the template. This is probably not exactly true in flint. The wear of the rolls, and possibly carelessness in making the rolls originally, together with the ihct that the rails rolled in the same set of rolls were probably not all rolled at the same temperature, and consequently shrunk different amounts in cooling, might'each occasion small deviations from the height which the rail should have according to template. And yet Mr. Cloud, engineer of tests, has callipered the height of some 60 new rails here at Altoona during the year past, which rails were from different mills, and has found the variation from standard height not more than x ~ of an inch. It would seem, therefore, that the error arising from the assumption as to height of rail could not be large. The second assumption is that the original shape of the top of the head of the rails we are studying was that of the template. This is of course a question of accuracy in the original nmlmfaeture of the
5far., 1881.']
Dudley--Steel Rails.
185
rolls, and of their wear. :Now it is impossible to say how accurately the rolls were originally made, and how much they have worn out of shape when any raii was rolled. But in ally case the error must be very small, for any noteworthy deviation of the top of the head from smlldard wo'hld have given difficulty in securing a straight, even track, and would have caused t h e rejection of those rails by the inspector at the mills. The third assumption is that these rails have not suffered distortion as Co height by the service which they have endured. Upon this point I will say that I think an inspection of the diagrams in accompanying Plates 2, 3, 4 and 5 will convince any one that this distortion, if any, is excessively small. I t should be stated here that any deviation of ~hese rails, from template, such as beading over on one side of the head, or thicker or thinner webs than standard, or variation in the shape of the foot or in the under side of t h e head, can cause no error ; because in taking the areas with the planimeter, the diagrams from which the area of the worn section and the area of the original section were taken, coincide in every part except in that part of the original section which was traced-in fl'om template as above described. In other words, in getting the areas of the rails as rolled, the actual section of the rail just as it was rolled was used in every part, except that portion of the head which was golm from wear, which portion was restored from template. Fourth. H o w accurately can the areas be obtained by means of the planlmeter? The planimeter which was used gave readings to ]-~-~ of a square inch. If, therefore, the manipulation was such as to give the instrument its full chance, the maximum error in area could not be over , 10.0 of a square inch. As a matter of t~act, in taking tlle area~ with the planimeter, from three to five measuremeuts were made of each diagram, which measurement.s differed from each other not more than from one to three hundredths of an inch, and then a mean of these measurements was taken as representing the area of the diagram. I t seems probable, therefore, that the error arising from this cause could not in any case amount to more than one or two hundreths of a square inch. Now as to the weight per yard of the woru rails. How can this best be obtained? Obviously the most simple method would be to weigh each of the worn rails, and then divide the weight by the length. This was done with these rails, as has already been mentioned. But
186
Dudley--Steel Rails.
[Jour. Frank. Inst.,
when the results were obtained, and the original weight of the rails computed, by means of the areas as above described, it was found that some rails rolled at the same mill, at the same time, and with the same thickness of web and same shape of foot, differed from each other ill the original weight, as computed, from 1½ to 3 lbs. per~ard. The explanation of this seems to be that the half-inch sections before referred to, which were used in getting the areas, did not exactly represent the whole rail; in other words, the rail was unevenly worn. This explanation was confirmed by callipering the height of the halfinch sections, and comparing these heights with the heights of the rail in different places. I t became necessary, therefore, to devise some other means of obtaining the weight per yard of the worn rails, and it was finally decided to obtain the desired weights from the half-inch sections themselves before referred to. These sections were all, therefore, carefully weighed on a balance which weighed accurately to one-half a grain. If, now, the half-inch sections were exactly half an inch in thickness, a little consideration shows that it is possible in this way to get at the weight of a rail moreaccurately than it could be obtained by weighing the whole rail on a common platform scale, supposing, of course, that the rail was uniform in section throughout its length. For there are 72 half-inch sections in a yard, and in a 30-foot rail 720 half-inch sections. If, now, the error in weighing a half-inch section is one-half a grain, the total error in the weight of. a 30-foot rail would be 720 half grains, or 360 grains, which is a little less than an ounce. Inasmuch, therefore, as ordinary platform scales do not generally weigh closer than half a pound, it is evident that, so far as accuracy of weighing is concerned, the method of getting at the weight per yard of these worn rails, by weighing the half-inch sections, leaves nothing to be desired. But were these half-inch sections exactly half an inch thick ? An examination of these sections by means of vernier callipers, reading to one-thousandth of an inch, showed that they were not exactly half an inch thick, and that they varied in thickness in different parts of the section. The reasons for this seem to be, that it is almost impossible to set a tool so as to cut off" exactly half an inch from a rail, and that in cutting steel as hard as some of these rails were, the tool is apt to spring more or less, and thus give a section of varying thickness. It became necessary, therefore, to devise some method by means of' which the average thickness of these sections could be obtained. This was
.~[ar.,1881.]
Dudley--Steel Rails.
187
done as follows : The half-inch sections were all carefully weighed in distilled water, at a temperature of 66°F., on a balance weighing to half a grain. Tile difference between the weight so obtained and the weight of the same sections in ~tir gives the weight of a volume of water equal to the volume of the section. Now, this weight of water divided by the weight of a cubic inch of water, which was taken as 252"5 grains, gives the volume of space occupied by any section. This volmne being known, it is of course only necessary to divide the same by the area obtainc
188
Dudley--Steel Rails.
[Jour. Frank. Inst.,
nage, the chemical analyses, the weights per yard, and loss of metal, as well as the loss per million tons, and the complete results of the physical tests and the density, are given in tabular form in accompanying Plates 6 and 7. These tables will be discussed farther on. The density is the weight of a cubic inch of the steel. This weight was obtained by dividing the weight of each of the half-inch sections by its volume, the data for this purpose having been obtained as previously described. The figures given Under density are fractions of a pound. In looking over the original weights of the rails as obtained, it will doubtless be observed that, although these rails were all supposed to have weighed 67 lbs. or 56 lbs. per yard when rolled, the figures given differ both ways from these figures. These differences are principally to be accounted for by differences in the shape of the bases and differences in the thickness of the webs. With regard to the latter point, it is evident that of two rails rolled at the same mill, in dae same year, if one has a web three or five one-hundredths of an inch thicker than another--which means that the rolls were farther apart when the thick-web rail was rolledmthe thin-web rail will be lighter than the other. The thickness of the webs of all the rails in d~e series we are studying was carefully measured with vernier callipers, and the differences from the standard thickness were found to vary both ways, from nothing up to five one-hundreths of an inch thinner than standard, and seven one-hundredths thicker than standard. Finally, the differences in the density of the different rails still further helps us to account for the differences in weight per yard of the rails as rolled. The following is the history and tonnage of the different rails : No. 881. Steel of 1868. Was on tangent, north rail, north track, in Bennington Cut. In service from June, 1888, to July, 1879-11 yrs. 1 mo. Grade, 92"4 ft. to the mile. Tonnage, 55,546,811 tons,
No. 882. Steel of 1868. Was oil tangent, north rail, north track, in Bennington Cut. I n service from June, 1868, to July, 1879-11 yrs. 1 too. Grade, 92"4 ft. to the mile. Tonnage, 55,546,811 tons. No. 883. Steel of 1869. Was on tangent, north rail, north track, on Whip-poor-will Straight. In service from May, 1869, to July,
Mar., 1881.]
Dudley~Sleel Rails.
189
1879--10 yrs. 2 mos. Grade, 95"04 ft. to the mile. Tonnage, 52,174,969 tons. No. 884. Steel Of 1869. Was on tangent, north rail, north track, on Whip-poor-will Straight. In service from M:ay, 1869, to July, 1879--10 yrs. 2 mos. Grade, 95"04 ft. to the mile. Tonnage+ 52,174,969 tons. No. 885. Steel of 1868. Was on tangent, south rail, north track, just west of Allegrippus Station. In service from July, 1868, to July, 1879--11 yrs. Grade, 89"76 it. to the mile. Tonnage, 55,197,99~ tons. No. 886. Steel of 1868. Was on tangent, north rail, north track, just west of A!legrippus Station. In service from July, 1868, to July, 1879--11 yrs. Grade, 89"76 ft. to the mile. Tonnage, 55,197,994 tOlls. No. 887. Steel of 1868. Was on tangent, south rail, north track, west of Allegrippus Station. In service from July, 1868, to July,: 1879--11 years. Grade, 89'76 ft. to the mile. Tonnage, 55,197,994 tons. No. 888. Steel of 1868. Was on tangent, south rail, north track, west of Allegrippus Station. In service from July, 1868, to July, 1879--11 years. Grade, 89'76 ft. to the mile. Tonnage, 55,197,994 tons. No. 889. Steel of 1873. Was on tangent, south rail, south track, at South Fork. In service fl'om August, 1873, to July, 1 8 7 9 - 5 yrs. 11 mo~. Grade, 21'13 ft. to the mile. Tonnage 44,620,100 to~s. No. 890. Steel of 1873. Was on tangent, north rail, south track, at South Fork. In service from August, 1873, to July, 1879-5 yrs. 11 mos. Grade, 21"13 ft. to the mile. Tonnage, 44,620,100 tons. No. 891. Steel of 1872. Was oll tangent, south rail, south track, at Summer Hill water plug. In service from June, 1872, to July, 1879--7 yrs. 1 too. Gra.de, 40"13 ft. to the mile. Tonnage, 53,687,192 tons. No. 892. Steel of 1872. Was on tangent, south rail, south track, at Summer Hill water plug. In service from June, 1872, to July, 1879--7 yrs. 1 too. Grade, 40"13 ft. to the mile. Tonnage, 53,687,192 tons. No. 893. Steel of 1874. Was on tangent~ north rail, south track~
190
Dudley--Steel Rails.
[Jour. Frank. Inst.,
near Summer Hill. In service from May, 1874, to July, 1879-5 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tonnage, 38,088,572 tOlls. ~No. 894. Steel of 1874. Was on tangent, south rail, south track, east of Portage. In service from May, 1874, to July, 1879--5 yl~. 2 mos. Grade, 52"8 ft. to the mile. Tonnage, 38,088,572 tons. No. 895. Steel of 1874. Was on tangent, south rail, south track, east of Portage. In service from May, 1874, to July, 1879--5 yrs. 2 mos. Grade, 52"8 ft. to the mile. Tonnage, 38,088,572 tons. No. 896. Steel of 1874. Was on tangent, south rail, south track, east of Portage. Iu service from May, 1874, to July, 1879--5 yrs. 2 mos. Grade, 52"8 ft. to the mile. Tonnage, 38,088,572 tons. No. 897. Steel of 1868. Was on high side of 5 ° curve, north rail, north track, east of Bridge :No. 3, Summer Hill. In service from May, 1868, to July, 1879,--11 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tonnage, 52,370,617 tons. No. 898. Steel of 1868. Was on high side of 5 ° curve, north rail, north track, east of Bridge No: 3, Summer Hill. In service from May, 1868, to July, 1879--11 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tolmage, 52,370,617 tons. No. 899. Steel of 1868. Was on low side of 5 ° curve, south rail, north track~ east of bridge No. 3, Summer Hill. In service from May, 1868, to July, 1879--11 yrs. 2 mos. Grade, 21"12 i~. to the mile. Tommge, 52,370,617 tons. No. 900. Steel of 1868. Was on low side of 5 ° curve, north rail, north track, east of Bridge No. 3, Summer Hill. In service from May, 1868, to July, 1.879--11 yrs. 2 mos. Grade, 21"12 ft. to the mile. Tonnage, 52,370,617 tons. ~'o. 901. Steel of 1871. Was on low side of 5 ° curve, south rail, north track, in Gable's Cut, near Summer Hill. In service from August, 1871, to July, 1879--7 yrs. 11 inos. Grade, 39'6 ft. to the mile. Tonnage, 40,061,230 tons. No. 902. Steel of 1871. Was on high side of 5 ° curve, north rail, north track, in Gable's Cut, near Summer Hill. In service from August, 1871, to July, 1879--7 yrs. 11 mos. Grade, 39"6 ft. to the mile. Tonnage, 40,061,230 tons. No. 903. Steel of 1876. Was on high side of 5 ° curve, south rail, south track, west of Bridge No. 1, Summer Hill. In service from
Mar., 1881.]
Dudley--Steel Rails.
191
August, 1876, to July, 1879--2 yrs. 11 mos. Grade, 39"6 ft. to the mile. Tonnage, 21,504,824 tons. No. 904. Steel :of 1876. Was on low side of 5 ° curve, north rail, south track, west of bridge No. 1, near Summer Hill. In service from August, 1876, to July, 1879--2 yrs. 11 mos. Grade, 39"6 ft. to the mile. Tonnage, 21,504,824 tons. No. 905. Steel of 1868. Was on high side of 5 ° curve, north rail, north track, east end of Milmore Middle Siding. Ill service from June, 1868, to April, 1879--10 yrs. 10 mos. Grade, 47'52 ft. to the mile. Tonnage, 50,648,939 tons. ~o. 906. Steel of 1868. Was on low side of 5 ° curve, south rail, north track, east end of Wilmore Middle Siding. In service from June, 1868, to April, 1879--10 yrs. 10 mos. Grade, 47"52 ft. to the mile. Tonnage, 50,64S,939 tons. No. 907. Steel of 1874. Was on high side of 5 ° curve, north rail, north track, east of Wilmore Middle Siding. In service from June, 1874, to April, 1879--4 yrs. 10 mos. Grade, 47"52 ft. to the mile. Tonnage, 24,211,147 tons. No. 908. Steel of 1875. Was on low side of 4 ° curve, north rail, south track, near tower at east end of Ga]itzin Tunnel. In service . fi'om May, 1875, to July, 1879--4 yrs. 2 mos. Grade, 95"04 feet to the mile. Tonnage, 32,428,6]4 tons. No. 909. Steel o f 1875. Was on Mgh side of 4 ° curve, south rail, south track, near tower at east end of Galitzin Tunnel. In service from May, 1875, to July, 1879---4 yrs. 2 mos. Grade, 95"04 ft. to the mile. Tonnage, 32,428,614 tons. No. 910. Steel of 1871. Was on low side of 5 ° curve, south rail, south track, west of Allegrippus Station. In service from July, 1871, to July, 1879--8 yrs. Grade, 89"76 ft. to the mile. Tonnage, 62,813,664 tons. No. 911. Steel of 1877. Was on low side of 5 ° curve, south rail, south track, second curve west of Allegrippus Tower. ]n service from April, 1877, to July, 1879--2 yrs. 3 mos. Grade, 89"76 i't. to the mile. Tolmage, 17,226,993 tons. No. 912. Steel of 1873. Was on high side of 5 ° curwi north rail, south track, second curve west of Allegrippus Tower. In service from July, 1873, to July, 1879--6 yrs. Grade , 89"76 ft. to the mile. Tonnage, 47,438,145 tons. No. 913. Steel of 1870. Was on tangent, north rail, north track,
192
Dudley--Steel Rails,
[Jour. Frank. Inst.,
600 feet west of mile post 211 from Philadelphia. In service from March, 1870, to July, 1879--9 yrs. 4 mos. On level. Tonnage, 45,855,101 tons. No. 914. Steel of -1870. Was on tangent, south rail, north track, 600 feet west of mile post 211 from Philadelphia. In service from March, 1870, to July, 1879--9 yrs. 4 mos. On level. Tonnage, 45,855,101 tons. No. 915. Steel of 1873. Was o n tangent, south rail, north track, east of Petersburg Toolhouse. In service from June, 1873, to July, 1879--6 yrs. 1 too. On level. Tonnage, 31,514,889 tons. No. 916. Steel of 1873. Was on tangent, north rail, north track, just west of Ardenheim. In Service from July, 1873, to July, 1879 - - 6 yrs. On level. Tonnage, 31,127,829 tons. No. 917. Steel of 1873. Was on tangent, south rail, north track, just west of Ardenheim. In service from July, 1873, to July, 1879 --6 yrs. On level. Tonnage, 31,127,829 tons. No. 918. Steel of 1869. Was on tangent, north rail, north track, west of Vandevaner's Bridge. I l l service from January, 1869, to July, 1879--10 yrs. 6 mos. On level. Tonnage, 51,720,011 tons. No. 919. Steel of 1869. Was on tangent, north rail, north track, west of Vandevaner's Bridge. In service from January, 1869, to July, 1879--10 yrs: 6 mos. On level. Tonnage, 51,720,011 tons. No. 920. Steel of 1867. Was on tangent, south rail, north track, at Jackstown Water-trough. In service fi'om December, 1867, to July, 1879--11 yrs. 7 mos. On level. Tonnage, 52,991,684 tons. No. 921. Steel of ]874. Was on tangent, north rail, north track, 200 feet east of mile post 211 from Philadelphia. In service from March, 1874, to July, 1879--5 yrs. 4 mos. On level. Tonnage, 27,622,230 tons. .No. 922. Steel of 1874. Was on tangent, south rail, north track, 200 ft. east of mile post 211 from Philadelphia. In service from March, 1874, to July, 1879--5 yrs. 4 mos. On level. Tonnage, 27,622,230 tons. No. 923. Steel of 1873. Was on tangent, south rail, north track, east of Petersburg Toolhouse. In service from June, 1873, to July, 1879--6 yrs. 1 mo. On level. Tonnage, 31,514,889 tons. No. 924. Steel of 1875. Was on taugent, south rail, south track, west of Vandevaner's Bridge. Iu service from June, 1875, to July, 1879--4 yrs. 1 mo. On level. Tounage, 36,349,989 tons.
Mar., 1881.]
Dudley--Steel Rait~.
193
No. 925. Steel of 1875. Was on tangent, south rail, south track, west of Vandevaner's Bridge. 1,1 service from June, 1875, to July, 1879--4 yrs. 1 too. On level. Tonnage, 36,349,989 tons. No. 926. Steel of 1870. Was on tangent, south rail, south track, 1500 ~eet west of mile post 182 from Philadelphia. In service from April, 1870, to July, 1879--9 yrs. 3 mos. On level. Tonnage, 76,409,123 tons. No. 927. Steel of 1870. Was on tangent, south rail, south track, 1500 feet west of mile post 182 from Philadell~fia. In service from April, 1870, to July, 1879--9 yrs. 3 mos. On level. Tonnage, 76,409,123 tons. No. 928. Steel of 1874. Was on tangent, south rail, south track, 200 feet east of mile post 181 from Philadelphia. In service fl'om JuDy, 1874, to July, 1879--5 yrs. On level. Tonnage, 43,610,150 tons. No. 929. Steel of 1867. Was on high side of 2¼° curve, north rail~ north track, 250 feet east of mile post 194 from Philadelphia. In service from December: 1867, to July, 1879--11 yrs. 7 mos. On level. Tonnage, 52,991,684 tons. No. 9:30. Steel of 1867. Was on low side of 2} ° curve, south rail, north track, 250 feet east of mile post 194 from Philadelphia. In service from December, 1867, to July, 1879--11 yrs. 7 mos. On level. Tonnage, 52,991,684 tons. No. 931. Steel of" 1868. Was on high side of 3 ° curve, south rail, north track, 400 feet east of mile post 191 from Philadelphia. In service from January, 1869, to July, 1879--10 yrs. 6 mos. On level. Tonnage, 51,720,011 tons. No. 932. Steel of 1868. Was on low side of 3 ° curve, north rail, north track, 400 feet east of mile post 191 from Philadelphia. In service from January, 1869, to July, 1879--10 yrs. 6 mos. On level. Tonnage, 51,720,011 tons. No. 933. Steel of 1869. Was on high side of 3} ° curve, north rail, south track, 1000 feet east of mile post 183 from Philadelphia. In service t~om November, 1869, to July, 1879--9 yrs. 8 mos. On level. Tonnage, 78,364,968 tons. No. 934. Steel of 1869. Was on low side of 3} ° curve, south rail, south track, 1000 i~ct east of mile post 183 from Philadelphia. In ~rvice from November, 1869, to July, 1879--9 yrs. 8 mos. On level. Tounagc, 78,364,968 tons. WHOLE:No.VoL. CXI.--(THIR]) SERI~S.Vol. lxxxi.) 13
194
Dudley--JS'ieel Rails.
[Jour. Frank.Inst.,
:No. 935. Steel of 1866. Was on high side of 2 ° curve, north rail, south track, at McVeytown Station. In service from September, 1866, to June, 1879~12 yrs. 9 mos. On level. Tonnage, 92,025,478 tons. :No. 936. Steel of 1866. Was on low side of 2 ° curve, soutll rail, south track, at McVeytown Station. In service from September, 1866, to June, 1879--12 yrs. 9 mos. On level. Tonnage, 92,025,478 tt)n No. 937. Steel of 1~876. Was on low side of 2 ° curve, south rail, south track, 200 feet west of mile post 192 from Philadelphia. In service from March, 1876, to July, 1879--3 yrs. 4 mos. On level. Tonnage, 29,905,122 tons. No. 938. Steel of 1876. Was on high side of 2 ° curve, north rail, south track, 200 feet west of mile post 192 from Philadelphia. In service from March, 1876, to July, 1879--3 yrs. 4 mos. On level. Tonnage, 29,905,122 tons. :No. 939. Steel of 1875. Was on the low side of 4 ° curve, south rail, south track, 250 feet east of Mount Union Bridge. In service from May, 1875, to July, 1879--4 yrs. 2 mos. On level. Tonnage, 37,150,179 tons. No. 940. Steel of 1875. Was on the high side of 4 ° curve, north rail, south track, 250 feet cast of Mount Ullion Bridge. In service from May, 1s75, to July, 187.9--4 yrs. 2 mos. On level. Tonnage, 37,150,179 tons. No. 941. Steel of 1874. Was on high side of 3 ° curve, south rail, south track, 300 feet east of Manayunk. In service from September, 1874 to July, 1879--4 yrs. 10 mos. On level. Tonnage, 42,277,638 tons. No. 942. Steel of 1874. Was on low side of 3 ° curve, north rail, south track, 300 feet east of Manayunk. In service from September, 1874, to July, 1 8 7 9 - 4 yrs. 10 mos. On level. Tonnage, 42,277,638 tons. No. 943. Steel of 1873. Was on the low side of 2 ° curve, south rail, south track, 2000 feet west of mile post 179 t~'om Philadelphia. Ill service from April, 1873, to July, 1879--6 yrs. 3 mos. On level. Tonnage, 55,127,464 tolls. No. 944. Steel of 1873. Was oil high side of 2 ° curve, north rail. south track, 2000 feet west of mile "post 179 from Philadelphia. In
Mar., 1881.]
.Dudley--Steel Rails.
195
service from April, 1873, to July, 1879--6 yrs. 3 mos. On level. Tonnage, 55,127,464 tons. Accompanying Plates 6 and 7 give in tal)ular form all the essential data of the history of the rails, together with the chemical analyses, weights, and results of the chemical tests, as has been before stated. To a study of these tables attention is now directed. As will be observed, the rails have been arranged in six groups, according to the kind of service to which they have been subjected. The first group of 16 rails did their service on grade tangents; the second group of 8 rails on the low side of grade curves; the third group of 8 rails on the high side of grade curves; the fourth group of 16 rails on level tangents ; the fifth group of 8 rails on the low side of level curves ; and the sixth group of 8 rails on the high side of level curves. As will likewise be observed, the rails within the groups have been arranged in regular order according to loss of' metal per million tons, the rail having lost least metal per million tolls being placed first. The first rails in the several groups are therefore the slower wearing, while the latter rails are the more rapid wearing. This arrangement enables us to compare the slower-wearing rails of the different kinds of service with the more rapid-wearing ones, and to discover what physical qualities, and what chemical composition, are characteristic of the slower-wearing as well as the faster-wearing rails. Giving our attention now to the tables, I think the first observation will be that there is no absolute gradation in physical qualities, or in chenfical composition, applying to every rail in each group, which corresponds to the gradation in amount of metal lost per million tolls. In other words, if we take tensile strength, or elongation in the physical qualities, or the carbon or phosphorus in the chemical composition, we do not find that the first rail in each group is characterized by a certain figure in any one or nmre of these respects, while the last rail in the group is characterized by another different figure, and all the intermediate rails arrange themselves in respect to any of these peculiarities, between these two extreme rails of the group. Nor do I think we ought to expect such uniibrmity, aud for the ibllowing reasons : First. Whatever errors there may have been in determining the loss of metal of these rails or in the tonnages, as has been before described, will of course have an influence in determining the position
196
Dudley--Steel Rails.
[Jour. Frank. Inst.~
of any rail in the group to which it belongs. And, as has already been said, while I do not think that these errors, whatever they may be, are sufficient to seriously obscure or counteract the general conclusions which the results are calculated to teach, yet these errors may possibly be large enough to give some rail a position in the groul~ which it should not occupy. Second. I am not aware that it is known as yet exactly what wear is, or what it is dependent upon. Some considerations in regard t~ the nature of wear will be advanced further on ; but whether wear is a direct function of the tensile strength of steel, or of its elongation~ or of its elastic limit, or of its resilience, or any combination of these, or indeed, as seems somewhat probable, of the amount of distortion by bending that a piece of steel will suffer, is a problem yet to be solved. I t may be that wear bears a direct ratio to the elongation within the elastic linfit, or to the amount of work done within the elastic limit whenever a particle is worn off; or, indeed, wear may be a direct function of the granular structure of steel, the wear being more rapid as the granular structure is coarser, or vice versa. Or, finally, wear may be due to a combinations of causes, and each cause may require for its elucidation a different physical test. I t is perhaps, therefore, not strange, in view of' this want of knowledge, that, in a series of steels arranged according to h)ss of metal, or on the principle of wear, the physical qualities which we are now able to measure should not show uniform gradations throughout the series. But, it seems t~ me, this reasoning does not tend to throw doubt on our ability to draw conclusions that will be valuable from the work which we have in hand. :For the question we are studying is not what does wear depend upon, but what chemical composition, and what physical properties, are, in general, characteristic of such rails as have in actual service given least loss of metal per million tons' burden ? While the answer to this question may not solve the whole problem of wear, yet it cannot fail to throw light on the relation between the chemistry and physics of steel and its wearing power, which, after all, with our present knowledge of steel metallurgy, is all the knowledge we are able to utilize. (To be continued.)